Increased breast cancer risk among sisters of breast cancer cases, especially when the proband was young at diagnosis, suggests that breast cancer is in part heritable. However, low concordance in identical twins points to a role for exposures or gene-&#8208; environment interactions. Young-&#8208;onset breast cancer is so uncommon that few studies have been large enough to evaluate its specific causes; nevertheless, some risk factors are known to differ between young-&#8208; and older-&#8208;onset disease. For example, overweight women have increased risk after menopause, but reduced risk before menopause, suggesting distinct mechanisms. Genome-&#8208;wide association studies for breast cancer have had limited success, but analyses have typically considered one single-&#8208;nucleotide polymorphism (SNP) at a time, neglecting interactions with nongenetic factors and epistasis. Case-&#8208;control studies cannot probe possible parent-&#8208;of-&#8208;origin effects or find maternal genetic variants that can prenatally affect the environment of the developing female fetus in a way that influences her later risk. Only family-&#8208;based studies permit such discoveries. We recently carried out the family-&#8208;based Two Sister Study with funding from Susan G. Komen for the Cure. We enrolled families where a case daughter was diagnosed recently and under age 50, and a sister who had not had breast cancer was a member of the Sister Study cohort. We collected DNA and the same extensive exposure data from both sisters, as well as DNA from their parents when possible. In the proposed project, we would use the Illumina 700K HumanOmni +250K Exome Array chip, augmented by imputations based on the Thousand Genome Project, to genotype samples from 1400 families with a young-&#8208;onset breast cancer case, and use family-&#8208;based statistical methods to assess the combined roles of multiple genes and nongenetic cofactors. This comprehensive characterizing of the genome would allow us to identify now-&#8208;unsuspected genetic variants, including haplotypes, that may be specific to young-&#8208;onset breast cancer. Preliminary data suggest that parents may be important: family history data from the Sister Study reveal that in families where a sister was diagnosed before age 50 many more maternal grandmothers than paternal grandmothers had breast cancer. This difference in the grandmothers suggests a mechanism through the maternal line, perhaps either a prenatal effect of maternal genes on the developing daughter fetus, or a parent-&#8208;of-&#8208;origin effect where only the copy inherited from the mother is expressed. The parental DNA we collected would allow us to probe such effects. Broad assessment of the genomes of these families will also facilitate future studies of genetic factors that contribute to treatment outcomes and long recurrence-&#8208;free survival. Breast cancer in young women may not always be preventable, but it could one day become universally survivable.